Previous day | Jump to hour: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | Next day

Seconds: Show Hide | Joins: Show Hide | View raw
Font: Serif Sans-Serif Monospace | Size: Small Medium Large

Click in the nick column to highlight everything a person has said.
The Logo icon identifies that the person is a core developer (has commit access).

Notice: Only Gecko based browsers prior to FF4 support the multipart/mixed "server push" method used by this log reader to auto-update. Since you do not appear to use such a browser, this page will simply show the current log, and not automatically update.

#rockbox log for 2011-01-07

00:04:17 Quit GeekShadow (Ping timeout: 240 seconds)
00:05:16*TheSeven wonders why nobody seems to ever call audiohw_set_master_vol
00:08:14*TheSeven glares at firmware/sound.c line 252 and falls off his chair
00:09:45n1skugel: care to elaborate?
00:09:55kugeln1s: download fails
00:10:14n1swhat are you building?
00:10:15kugelI executed it and selected "s m e i", and the download fails
00:10:39n1swhich download?
00:11:32n1shmm, yes atypo seems to have crept in
00:12:07n1sor something
00:13:26 Quit benedikt93|AFK (Quit: Bye ;))
00:13:41 Quit Judas_PhD (Quit: This is a quitting message)
00:15:03 Join Judas_PhD [0] (
00:15:09 Join Rob2222 [0] (
00:15:51n1soh, the geolocated mirror's seem to have different dir layout
00:19:24 Join JdGordon1 [0] (
00:20:45n1skugel: if you set GNU_MIRROR= when running the script it works
00:21:02n1si wonder why they can't use the same dir structure :(
00:21:26 Quit Judas_PhD (Quit: This is a quitting message)
00:21:33n1sthe nice thing with the mirrors on the gcc site is that they have all the things we need
00:24:23 Join Judas_PhD [0] (
00:24:23 Quit lc0806607 (Read error: Connection reset by peer)
00:25:38n1sso we either hardcode one of the gcc mirrors, or hardcode the url only for mpc
00:26:02 Quit Judas_PhD (Client Quit)
00:26:05***Saving seen data "./dancer.seen"
00:26:17n1ssince there appears to be no nice generic geolocation aware url for gcc mirrors and the regular gnu mirrors don't have mpc
00:26:48CIA-7New commit by 03theseven (r28980): Fix volume control on iPod Classic
00:28:10 Join Judas_PhD [0] (
00:29:27CIA-7r28980 build result: All green
00:31:10*TheSeven wonders how to figure out why playback needs so much processing power
00:32:59saratogahow much does it need?
00:33:19TheSevenit doesn't even play back WAV files (containing cd-quality PCM) without stuttering
00:33:33TheSevenFLAC, however, works nicely (but seems to need almost 200MHz)
00:33:52saratogaWAV is faster then flac so its probably not the codec itself
00:33:57saratogadisk driver maybe?
00:34:19TheSeventhat fun goes on after it completes buffering, so it can't be the disk driver
00:34:33TheSevenMP3 stutters as well
00:34:36saratogahow big is your audio buffer?
00:34:48saratogacompressed audio?
00:35:01saratogaand do you have cache enabled?
00:35:11 Quit Rob2222 (Quit: Rob2222)
00:35:21TheSeventhe PCM buffer is ~500KB, the compressed audio buffer ~60MB
00:35:27TheSevenand yes, the caches are enabled
00:35:35TheSeventhe test_codec results look sane
00:35:43saratogahow fast for MP3?
00:35:48TheSeven(WAV needing 2.5MHz, FLAC like 20MHz, MP3 like 45MHz)
00:36:03saratogathis is on arm9E?
00:36:14TheSevenyes, running at 216 MHz as far as i can tell
00:36:25saratogathats quite slow compared to AMS
00:36:34TheSeven(some quick measurements showed ~190MHz, so that should be right)
00:36:49TheSevenmemory bandwidth is >100MB/s, so that's not an issue either
00:36:58*TheSeven thinks the only thing left is the PCM buffer
00:37:22TheSevenPCM driver*
00:37:30saratogatest_codec doesn't yeild IIRC
00:37:36saratogaplayback does
00:37:59saratogamaybe some issue with scheduling?
00:38:35TheSevenbuffering seems to be over-yielding (in the ATA driver), but apart from that I can't see what might be going on
00:38:41saratogaif you open up mpa.c or some similar codec and make it yield less often does the stuttering get better?
00:38:49TheSevenisn't it basically only the codec thread that's eating CPU power once buffering has finished?
00:39:06saratogaassuming the screen is off i think so
00:39:16 Quit mystica555 (Ping timeout: 240 seconds)
00:39:35saratogaalthough i guess its possible you could have something busy waiting in a driver or something by mistake
00:39:35TheSevenWPS or sitting in the main menu (cabbiev2) doesn't make a difference
00:39:42froggymanis there any reason why volume control on the fuze seems to take a while to initiate? The new volume that the person sets doesn't seem to take effect till a couple seconds after you change it
00:39:49TheSeventhe UI is almost unresponsive
00:39:56TheSeven(except for FLAC)
00:40:08saratogayeah that'll happen if the PCM buffer gets low
00:40:20saratogait'll starve the GUI first
00:40:22 Quit Strife89 (Read error: Connection reset by peer)
00:41:11 Join Strife89 [0] (
00:41:59saratogatest codec fakes most of the playback and buffering APIs, so its not unreasonable that it could be different, but i'm not sure how this could be a hardware specific issue
00:42:50TheSevenfor FLAC, the PCM buffer is always around half full full
00:42:57saratogaif you disable the audio driver (just have it discard samples) does that improve things?
00:43:14TheSevenand buffering and the codec seem to be fighting badly while buffering is still active
00:43:24TheSevenbuffering almost can't keep up with playback
00:43:24saratogai wonder if you're just getting spammed with interrupt requests
00:43:54TheSevenbut why does FLAC work if WAV doesn't?
00:44:05TheSevenare the PCM chunks insanely small for WAV?
00:44:28saratogai can't remember how libpcm does it
00:44:33saratogabut they may be smaller
00:44:41saratogaa lot of codecs force larger PCM chunks then we'd like
00:44:44TheSeventhe PCM driver can handle up to ~9000 samples (which is the highest number i've seen so far in a single chunk) with a single IRQ
00:44:45saratogasince we want to yield often
00:45:29saratogamost codecs generate about 2048 samples at a time
00:45:36*TheSeven means 9000 samples per channel, so 18000 samples or 36KB total
00:45:42saratogawhich codec?
00:46:04TheSeveni think that was MP3 (which was the first one i was testing with)
00:46:27saratogaok then thats the DSP engine combining multiple frames
00:49:50 Join Rob2222 [0] (
00:51:03TheSevenmust be some weird bug in my pcm driver...
00:51:19CIA-7New commit by 03nls (r28981): Hardcode gcc mirror for now as the gnu mirrors have a different dir layout and are missing the infrastructure dir which was the whole ...
00:53:53CIA-7r28981 build result: All green
00:55:16 Quit factor (Ping timeout: 240 seconds)
00:55:41gevaertsn1s: at what point should we start transitioning build clients?
00:56:18 Join notdavid [0] (
00:56:39 Join factor [0] (~factor@
00:56:46n1sgevaerts: i don't know how the build clients stuff work really so how would we do that?
00:56:53 Quit chattr (Quit: gone)
00:57:11gevaertsIt's a bit annoying that the tools have exactly the same name
00:57:26 Quit TheLemonMan (Quit: free(me))
00:57:28 Quit panni_ (Ping timeout: 246 seconds)
00:58:11n1sgevaerts: does the build master have a way to determine if the client uses the correct version of something?
00:58:52kugelwe just need a toolchain test which is a little more sophisticated than which
00:59:01gevaertsIt just checks if it finds the compilers
01:00:07saratogaTheSeven: my wav files generate about 100 samples at a time, which is significantly less then most other formats I suspect
01:00:21kugelI don't think we need to worry though
01:00:44CIA-7New commit by 03theseven (r28982): iPod Classic: Unmute amplifiers when setting volume (fixes resuming paused playback)
01:00:46n1sunless we need it to be super smooth i'd say 1) commit configure/ changes and send mail to client owners 2) disable all cf builds 3) reenable for people who have updated
01:01:02gevaerts2) isn't really needed
01:01:10kugelwe would probably add a new m68k-elf-gcc452 target (instead of only m68k) and trust people that have it in their
01:01:13gevaertsWe can do a different architecture string in the clients
01:01:22gevaerts(as kugel said)
01:01:52n1swouldn't the only difference be a different error?
01:01:53CIA-7r28982 build result: All green
01:01:57gevaertsAnd yes, as far as I remember the plan is to do that
01:02:51gevaertsn1s: no. If a build client announces "m68k", it won't get builds that require "m68k-elf-gcc452"
01:03:47gevaertsThe *only* problem is that this list of announced architectures is hand-built by the build client maintainers, so there's some room for human error there
01:04:22n1sok, so should we keep the option to build the old toolchain in ?
01:05:57gevaertsI think we should somehow have the clients report compiler (and binutils?) versions to the server automatically
01:06:17n1syes, that would be good
01:06:29gevaertsSo when the client admin specifies "m68k-elf-gcc", the script expands that to m68k-elf-gcc452 or m68k-elf-gcc346
01:07:02n1sand the server should be able to decide on version requirements per-build as a sim build for example doesn't have such stric requirements
01:07:31 Join panni_ [0] (
01:07:32n1sanyway, i'll look into adding the suffix to the script tomorrow
01:07:35gevaertsThat does complicate things a bit...
01:07:48n1sthen we can probably start the transition
01:07:52JdGordon1isnt the sim a completly different beast in the build system?
01:08:04gevaertshm, yes, it is
01:08:14gevaertsWell, yes and no
01:08:25gevaertsn1s: I meant doing that in the build client script
01:08:33gevaertsNot necessarily in
01:08:55n1sgevaerts: ok, then i'm not really following :)
01:09:00kugeln1s: i don't think we need the old toolchain around
01:09:03n1swhat do i need to do?
01:09:13kugelnot if it's so complicated anyway
01:09:49 Join madalu [0] (~user@unaffiliated/madalu)
01:09:52gevaertsn1s: can you handle perl?
01:10:03 Quit MethoS- (Remote host closed the connection)
01:10:19n1skugel: i'd like to get rid of it too, only someone should keep one to do an eventual 3.7.2 release
01:10:26kugelwe only kept the old arm toolchain because it was a) easy and b) not all targets where tested when the transition happened
01:10:46n1sgevaerts: well, i'm not very used to it but have hacked a few of our lang scripts
01:10:47gevaertsright. We don't want to change toolchains for the stable branch
01:11:25gevaertsBut the stable branch will still have its own, so it's always easy to build a toolchain for it
01:11:54gevaertsn1s: I think we should wait for Zagor to appear before doing this. He knows how these scripts work :)
01:12:17n1syeah, any idea when he'll be back?
01:12:31gevaertsno idea. Maybe B4gder knows?
01:12:35 Join MethoS- [0] (~clemens@
01:14:03 Quit pamaury (Remote host closed the connection)
01:15:09 Join eRivas [0] (
01:15:20gevaertsn1s: anyway, the script has a bit where it checks that every arch specified with -archlist has a matching compiler. That check could presumably expand the arch string to append versions
01:16:16gevaertshm, actually...
01:17:36gevaertsIt now does "which m68k-elf-gcc", but somehow the gcc build also produces "m68k-elf-gcc-3.4.6" (and similar for the others). That means that the client script could check for the right compiler version (not binutils unfortunately, but that's probably minor)
01:17:51gevaertsSo all we have to do is update and push a new version
01:18:21gevaertsPushing a new version has to be done by a server admin, so we need one of the (other :) Swedes for that
01:19:30n1syes, clever. While i might be able to do that i'm not comfortable coding things i can't test :)
01:20:16gevaertsAt the same time, when we transition to 4.5, we change the announced string to also have the version (like it's done for eabi). That way (as long as there's only one toolchain in the path), there's a cross-check between what the client admin configures and what's actually there
01:20:27 Join krazykit [0] (
01:21:49n1svery simple, i like
01:22:11 Join kkit|sh [0] (
01:22:17 Quit kkit|sh (Client Quit)
01:22:31 Join kkit|sh [0] (
01:24:21gevaertsok. So we wait for Zagor, and then we should be able to switch within a day or so
01:24:42gevaertsIt might take longer before all build clients are updated, but that's not a huge problem
01:25:23n1ssounds great
01:25:29gevaertsAs long as we have a few when we switch (but not all yet, otherwise unexpected commits from just before the switch wouldn't build), we're fine
01:28:06 Quit liar (Quit: Leaving)
01:30:18 Join chattr [0] (
01:33:42 Quit kugel (Remote host closed the connection)
01:37:49eRivasis it possible to have a windows and a linux uisim sharing the same simdisk?
01:41:34 Quit FBI_Guy (Quit: I'm outta here)
01:42:41 Quit kadoban (Ping timeout: 240 seconds)
01:43:34 Quit Keripo (Quit: Leaving.)
01:46:26 Quit eRivas (Quit: CGI:IRC)
01:46:52 Join eRivas [0] (~je@
01:52:03[Saint]eRivas: Probably.
01:52:10[Saint]Why not just try it and see?
01:52:30 Join Str89|dormbox [0] (
01:53:04eRivasI did copy the .rockbox from my windows install and couldnt play songs in ubuntu, said missing codec or something like that
01:53:46[Saint]well, of course that won't work.
01:54:00[Saint]I thought you only meant for testing themes or something.
01:54:04n1sno, codecs and plugins are still binaries built for whatever system
01:55:05 Quit n1s (Quit: Lämnar)
01:56:19eRivasits just I wont be able to take screenshots of the wps with the message popping up, but yeah, menus and stuff work fine
01:56:34 Quit saratoga (Quit: Page closed)
01:56:55[Saint]why do you *need* to take screenshots of it?
01:58:11[Saint]I don't understand what it is you're trying to do.
01:59:37eRivasno, right now its just for fun and convinience, since that would keep both sims 'synced', just wanted to get all the functionality
02:00:19eRivasand screenshots are for when I want to upload themes, but thats only happens once in the theme development, so really no that critical
02:00:41[Saint]what functionality are you missing?
02:01:05[Saint]And, I believe I have already told you how to capture screenshots on the device.
02:01:50eRivasyeah, screen dumping or f5 on the sim
02:02:11[Saint]I'm still curious about "just wanted to get all the functionality"
02:02:35[Saint]doesn't trying to run a windows binary in *nix give you *less* functionality ")
02:02:44[Saint]* :)
02:03:12eRivasoh, the linux one was built from svn, not a 'wined' version
02:04:57 Join Keripo [0] (
02:05:34 Join barrientos [0] (~barriento@
02:19:43 Join Strife1989 [0] (
02:22:10 Quit Str89|dormbox (Ping timeout: 246 seconds)
02:23:30 Join Str89|dormbox [0] (
02:24:19 Quit simon_ (Quit: leaving)
02:25:30 Quit Strife1989 (Ping timeout: 250 seconds)
02:26:06***Saving seen data "./dancer.seen"
02:27:20 Join designate72 [0] (
02:29:03 Quit JesusFreak316 (Read error: Connection reset by peer)
02:30:55notdavidi just installed 3.7.1 on a nano 2g and have an issue where i must press one of the 5way navigation buttons before the scrollwheel is activated
02:31:15 Join BHSPitMonkey [0] (~stephen@unaffiliated/bhspitmonkey)
02:31:19notdavidvery annoying. are there any workarounds?
02:31:59 Join Strife1989 [0] (
02:32:31[Saint]You don't mean "all the time", do you?
02:32:48[Saint]I have that issue occasionally, and it is a known bug.
02:32:59notdavidyes, i mean ALL the time
02:33:17notdavidwell... if i keep it spinning it doesnt happen.
02:33:25[Saint]something must be wrong with your hardware.
02:34:16 Quit Str89|dormbox (Ping timeout: 240 seconds)
02:34:24[Saint]I have two Nano 2Nd Gens in front of me now running 3.7.1 (I just rolled back) with default settings, and I am not experiencing this.
02:34:58 Join simonrvn_ [0] (simon@
02:34:59[Saint]Could you try backing up your config.cfg and restoring the default setting state?
02:35:06notdavidi did just do a "repair" on it so maybe youre right
02:35:23notdavidi was just about to do this and reinstall the whole thing just incase
02:37:06[Saint]I would like to verify first that it is not an issue with your configuration file, if you'd like to back it up (you can just drag it off your player onto the desktop or something for this test) and try with default settings on the player it would help us to know if it is an issue with the hardware or not.
02:37:55 Quit simonrvn (Ping timeout: 240 seconds)
02:37:56 Nick simonrvn_ is now known as simonrvn (simon@
02:38:17 Quit barrientos (Quit: Saliendo)
02:40:07[Saint]notdavid: ^ ?
02:42:03 Quit Strife1989 (Quit: Leaving)
02:43:14notdavidi deleted the cfg an still persists.
02:43:37notdavidwhere is the option in rb to dump settings?
02:47:23 Join Strife89DM [0] (
02:47:31[Saint]Settings −−> Manage Settings −−> Save .cfg File
02:47:37[Saint]na: ^
02:47:53[Saint]*notdavid: ^
02:48:52[Saint]notdavid: Can you test please, if the clickwheel behaves normally in the original Apple firmware?
02:48:53 Quit Judas_PhD (Quit: This is a quitting message)
02:49:21[Saint]If it does not function correctly there, it is definitely a hardware problem.
02:51:03notdavid[Saint]: yup, wheel worked fine before i loaded rb onto it
02:51:48[Saint]can you please test the Apple firmware though to see if it still functions correctly.
02:53:16[Saint]I have 6 Nano 2nd gens, and I only ever see this bug every once in a while and it is known. ALL the time though very strongly points to something else.
02:53:25 Join fdinel [0] (
02:53:56notdavidwill do. give me a lil while though
02:56:35 Join Judas_PhD [0] (
02:59:15 Join Strife1989 [0] (
03:01:30notdavid[Saint]: rbutil refuses to uninstall the bootloader
03:02:08[Saint]this sounds like disk corruption, possibly.
03:02:20 Quit Strife89DM (Ping timeout: 250 seconds)
03:02:23[Saint]restoring with iTunes will uninstall the bootloader.
03:03:40notdavidany other way besides itunes? wine still doesnt have proper usb support
03:03:53[Saint]boot the player into disk mode, and iTunes should have no problems recognising it if you want to try using your operating system to check for disk corruption.
03:04:08[Saint]dammit, why did I say iTunes there.
03:04:37[Saint]I'm not familiar with *nix disk utilities sorry.
03:06:42[Saint]notdavid: There is this:
03:07:07[Saint]that's all the documentation I know of for restoring an iPod without iTunes
03:12:26notdavid[Saint]: the ipod shows a white screen of death telling me to use itune. i should have consulted the wiki first. on the plus side, i removed rb :)
03:13:50[Saint]On a side note, I should mention that if you have access to a windows based PC it is highly reccommended that you use iTunes to restore the player.
03:15:38[Saint]If you have any questions about the iPod Manual Restore, *ask them!*, if you're unsure about anything with dd, don't blindly proceed. Incorrect use of dd can seriously harm your system.
03:22:43 Join simonrvn_ [0] (
03:24:55 Quit simonrvn (Disconnected by services)
03:24:56 Nick simonrvn_ is now known as simonrvn (
03:42:11notdavidoh to hell with this. ill just install react os and use itunes
03:42:47krazykitthere's the manual method on the IpodManualRestore wiki page (i think the name is)
03:44:54 Quit fdinel (Quit: Miranda IM! Smaller, Faster, Easier.
03:48:09 Quit BHSPitMonkey (Ping timeout: 264 seconds)
03:48:40notdavidkrazykit: i tried that but i still have the white screen of death
03:51:23 Quit madalu (Read error: Operation timed out)
03:53:41[Saint]iTunes is really the preferred method, as it "just works".
03:53:58[Saint]the only downside to it is, ....using iTunes. ;)
03:56:10notdavidim setting up react os in qemu right now so i can run itunes and have usb support
03:57:21 Join BHSPitMonkey [0] (~stephen@unaffiliated/bhspitmonkey)
04:03:21 Join fdinel [0] (
04:04:42 Quit designate72 (Ping timeout: 265 seconds)
04:14:19 Join Barahir_ [0] (
04:17:39 Quit Barahir (Ping timeout: 250 seconds)
04:17:51 Quit notdavid (Quit: quit)
04:24:06 Quit DerPapst (Quit: Leaving.)
04:26:09***Saving seen data "./dancer.seen"
04:31:36 Quit Judas_PhD (Quit: This is a quitting message)
04:34:25 Quit tchan (Read error: Connection reset by peer)
04:34:49 Join kadoban [0] (
04:35:15 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
04:35:33 Join Keripo1 [0] (
04:35:48 Quit Keripo (Ping timeout: 246 seconds)
04:40:05 Join Strife89DM [0] (
04:40:53 Quit Strife89DM (Client Quit)
04:41:58 Quit Strife1989 (Ping timeout: 260 seconds)
04:43:10 Quit pixelma (Disconnected by services)
04:43:12 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:43:14 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:44:25 Quit amiconn (Disconnected by services)
04:44:25 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:44:43 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:48:39 Quit MethoS- (Remote host closed the connection)
04:55:58 Quit TheSeven (Ping timeout: 260 seconds)
04:59:35 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
05:10:37 Quit Rob2222 (Ping timeout: 260 seconds)
05:10:38 Join Judas_PhD [0] (
05:29:56 Quit Horscht (Quit: Verlassend)
05:30:12 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
05:33:49 Quit JdGordon1 (Ping timeout: 240 seconds)
05:33:56 Join JdGordon1 [0] (
05:51:45 Quit fdinel (Quit: Miranda IM! Smaller, Faster, Easier.
05:54:18CIA-7New commit by 03rmenes (r28983): Reword the blind description for the MPIO HD200 manual ...
05:56:33CIA-7r28983 build result: All green
06:01:57 Quit eRivas (Quit: Saliendo)
06:05:27 Quit parafin (Ping timeout: 260 seconds)
06:06:31 Join parafin [0] (
06:17:31CIA-7New commit by 03jethead71 (r28984): MPEGPlyaer: A bit of audio mutation. Remove a useless thread state. Take some control over the buffer away from the audio thread itself. Some ...
06:18:37 Quit kadoban (Ping timeout: 255 seconds)
06:20:06CIA-7r28984 build result: All green
06:26:13***Saving seen data "./dancer.seen"
06:30:24CIA-7New commit by 03jethead71 (r28985): Put illustrative switch case back, because it clarifies intention...maybe ;).
06:32:45CIA-7r28985 build result: All green
07:02:02 Quit panni_ (Quit: ( :: NoNameScript 3.81 :: ))
07:06:11 Quit JdGordon1 (Quit: leaving)
07:07:45 Join kadoban [0] (
07:08:41 Quit factor (Ping timeout: 272 seconds)
07:15:47 Quit mikroflops (Quit: "foo")
07:16:07 Quit Keripo1 (Quit: Leaving.)
07:19:26 Quit rvvs89 (Ping timeout: 255 seconds)
07:20:08 Join rvvs89 [0] (
07:27:56 Quit pjm0616 (Quit: pjm0616)
07:29:33 Join T44 [0] (
07:31:21 Join mikroflops [0] (
07:33:02 Quit Topy (Ping timeout: 240 seconds)
07:42:11 Join kugel [0] (~kugel@
07:42:12 Quit kugel (Changing host)
07:42:12 Join kugel [0] (~kugel@rockbox/developer/kugel)
08:00:04 Quit BHSPitMonkey (Remote host closed the connection)
08:00:40 Quit rvvs89 (Remote host closed the connection)
08:21:20 Join Zarggg [0] (
08:21:45 Join Buschel [0] (
08:26:01 Quit Buschel (Client Quit)
08:26:14***Saving seen data "./dancer.seen"
08:27:42 Quit t0rc (Ping timeout: 250 seconds)
08:28:54 Quit kugel (Ping timeout: 260 seconds)
08:29:02 Join Topy [0] (
08:31:50 Quit T44 (Ping timeout: 240 seconds)
08:48:46 Join JuanArg [0] (
08:50:30JuanArgHello, I'm looking for help using tcctool with a Sansa c140, would anyone be willing to help?
08:53:58 Join kugel [0] (~kugel@rockbox/developer/kugel)
08:54:56JuanArgI already installed the driver to boot the device from USB and I read on the Cowon community that you need to add the device name to the command line, and I don't know how should I write it
08:55:52JuanArgI mean, should I write c100, c140, Sansa...? (Not spamming, just writing my question)
08:59:31 Quit JuanArg (Quit: CGI:IRC (EOF))
08:59:37 Join JuanArgPt2 [0] (
09:00:38 Nick JuanArgPt2 is now known as JuanArg (
09:01:22JuanArgMay I have forgotten any other piece of info?
09:04:05 Join pjm0616 [0] (~user@
09:04:42 Quit JdGordon| (Quit: leaving)
09:06:24[Saint]AlexP: Ping?
09:06:32 Join ender` [0] (
09:07:35[Saint]last time I checked JuanArg, you weren't AlexP :D
09:08:09linuxstbJuanArg: If you run tcctool without any parameters, it displays the list of device names.
09:08:49JuanArgSaint: Sorry, I missed the username you had typed before.
09:09:37[Saint]No problem, sorry I said "argh"...but I meant "Oh crap!", it's a coincidence it slightly matches your nick.
09:09:39[Saint]sorry ;)
09:09:41JuanArglinuxstb: thank you very much. I'll try that.
09:10:19JuanArgSaint: It's alright, no problem at all :)
09:14:46JuanArglinuxstb: I got it. Thank you very much again.
09:23:42 Join kevku [0] (~kevku@2001:7d0:0:f000::135d)
09:24:10 Join stoffel [0] (
09:26:18 Quit JuanArg (Quit: CGI:IRC)
09:29:37 Quit Judas_PhD (Quit: This is a quitting message)
09:29:40 Join petur [0] (d408b802@rockbox/developer/petur)
09:33:31 Quit [Saint] (Quit: I'm only going to Heaven if it feels like Hell, I'm only going to Heaven if it tastes like caramel...)
09:37:50 Join efyx [0] (
09:39:16 Quit kevku (Read error: Operation timed out)
09:39:43 Join kevku [0] (
09:47:08 Join Saij [0] (
10:08:17 Join Judas_PhD [0] (
10:21:34 Join [Saint] [0] (S_a_i_n_t@
10:26:09 Quit sasquatch (Quit: translocating)
10:26:17***Saving seen data "./dancer.seen"
10:31:02 Join tails__ [0] (
10:32:52 Quit Saij (Read error: Connection reset by peer)
11:05:55 Join DerPapst [0] (
11:08:32 Quit Judas_PhD (Quit: This is a quitting message)
11:09:33 Quit kevku (Remote host closed the connection)
11:14:39 Join Saij [0] (
11:14:41 Part Saij
11:14:49 Join alex121997 [0] (
11:15:33alex121997how to get MIDI playing on rockbox?
11:16:16alex121997Or better said, how to get MID files playing on rockbox?
11:16:27alex121997Thank you.
11:18:26 Join n1s [0] (
11:18:34 Quit n1s (Changing host)
11:18:34 Join n1s [0] (~n1s@rockbox/developer/n1s)
11:19:00pixelmawhich player do you own and what's our manual telling you?
11:19:47alex121997I own an iPod touch 5g (video) 30 gb.
11:20:10alex121997I gonna read the manual...
11:20:19alex121997sec please.
11:21:03 Join Saij [0] (
11:21:40 Quit kadoban (Ping timeout: 265 seconds)
11:23:56 Quit kugel (Ping timeout: 272 seconds)
11:27:53 Join kevku [0] (~kevku@2001:7d0:0:f000::135d)
11:31:32 Quit stoffel (Ping timeout: 272 seconds)
11:36:35 Part Saij
11:39:50 Quit efyx (Quit: Quitte)
11:41:14 Quit alex121997 (Ping timeout: 240 seconds)
11:42:14 Join Saij [0] (
11:47:38 Quit Saij (Read error: Connection reset by peer)
12:00:49 Join max131 [0] (
12:01:48 Quit petur (Quit: Page closed)
12:02:13max131damnit, i've got a file system corruption again on the fuze-v1. I'm running r28979 now, but i had a 'delayed write failed' error in windows xp. So the USB stack is not as solid as i expected??
12:03:09max131this time around i was moving nearly 1G of files from the hard disk to the internal memory of the fuze, _not_ to the uSD
12:04:35max131after the delayed write failed, the fuze locks up entirely, the only thing i can do is a 'hard reset', i.e. keeping the slider on the right of the fuze up for more than 10 seconds
12:05:04n1sthat sounds just like the recent sd issues
12:05:43n1sthe revert in r28925 made them go away for me at least
12:06:06n1smax131: did it ever work right and can you reproduce this error?
12:06:12max131yeah, i thought that too
12:06:38max131reproducability is not that easy
12:07:03max131i thought things got solved with r28295, that's why i did some stress tests during past days
12:07:31max131which all passed, where i did about 4 tests writing large amounts of data
12:07:53max131it just happens, sometimes ...
12:08:12max131if i find a way to reproduce, i'll let you peeps now
12:08:23n1sr28925 only reverted r28834 so if it didn't work before that there's no reason for it to work now
12:08:56 Join kugel [0] (~kugel@rockbox/developer/kugel)
12:09:06max131well, i thought the fs-corruptions started with r28834
12:09:44n1sthat's what i was trying to ask you before, did it ever work as in before that commit?
12:11:16max131i don't think i ever had fs-corruptions with pre-r28834 builds
12:13:03 Join stoffel [0] (
12:13:26max131and these fs-corruptions do occur on the internal storage
12:18:04 Join wodz [0] (
12:18:45wodzn1s: Have you conducted battery_bench tests gcc 3.4.6 vs. 4.5.2?
12:19:26max131back to work now, i'll keep you posted ...
12:19:57 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
12:20:06 Nick YPSY is now known as Ypsy (
12:20:12 Quit max131 (Quit: CGI:IRC (EOF))
12:21:20n1swodz: no
12:21:51wodzIt would be interesting
12:23:54n1sthey take forever and unless you use a codec that got significantly faster i wouldn't expect any notable change
12:25:18wodzIf you point me once more to the hd300 build with new gcc I may run battery_bench anyway
12:26:18***Saving seen data "./dancer.seen"
12:27:36n1sif you are interested in the codec speed changes,
12:28:32wodzwhat is final plan for switch? Post on ML and 'away we go'?
12:30:04n1swell, i had a discussion with gevaerts last night about how to do the distributed build system switch and he had a clever idea but we wanted to talk to Zagor about it first so we'll probably wait for him but after that is set up we are pretty much ready to go
12:30:32n1si need to retest one thing on cygwin
12:31:26n1sB4gder: any idea when Zagor will be back?
12:33:49gevaertsn1s: not just *wanted*. We need Zagor (or possibly B4gder, but he usually points to Zagor for these things) for the server-side work
12:40:03CIA-7New commit by 03rmenes (r28986): Add iPod Classic to the "Unusable" category. Also ...
12:40:21 Join LambdaCalculus37 [0] (
12:40:21 Quit LambdaCalculus37 (Changing host)
12:40:21 Join LambdaCalculus37 [0] (~rmenes@rockbox/staff/LambdaCalculus37)
12:41:15LambdaCalculus37Zagor, Bagder: (for the logs) Just made an update to index.t; can one of you guys update the front page?
12:42:50 Join bmbl [0] (~bmbl@unaffiliated/bmbl)
12:59:17n1sgevaerts: ah
13:02:08 Quit LambdaCalculus37 (Quit: Fwump)
13:03:37*TheSeven realizes that the µsec timer on the ipod classic is broken
13:04:48n1sTheSeven: is that brokeness the source of your problems?
13:04:58TheSevenno idea...
13:05:16TheSeveni don't think anything besides the clickwheel acceleration (and the profiling code i just added) uses it
13:06:10n1sah, so another timer is used for ticks?
13:07:16wodzWhat about system ticks? Are you sure it works as advertised? I had various strange effects on HD200 when tick was much longer due to bug in timer configuration
13:08:34TheSevencurrent_tick seems to advance properly at least
13:09:56CIA-7New commit by 03theseven (r28987): Fix iPod Classic USEC_TIMER
13:10:30TheSevenhm, if anything, the tick is too fast
13:10:56TheSevenat least my splashf(HZ, whatever) calls are only on the lcd for about half a second
13:11:34wodzso I guess this is the root of the problems with playback
13:12:12CIA-7r28987 build result: All green
13:12:51TheSevenhm, it's ticking ~60% too fast
13:15:54*kugel asked if the timer tick is correct a few days ago :)
13:18:38 Join Stephenc___ [0] (
13:20:12 Join MethoS- [0] (~clemens@
13:21:03 Quit kugel (Remote host closed the connection)
13:23:12Stephenc___anyone able to read todays irc log ?
13:24:21pixelma"Pressing the rocker in functions as..." - am I the only one who had to read this sentence a few times before understanding that the "in" belongs to the "pressing"? I think that the section got better overall but I had a bit of trouble with this tiny thing
13:25:34pixelmaStephenc___: I am (even the live reader) in firefox. Which browser do you use?
13:26:27CIA-7New commit by 03theseven (r28988): Fix iPod Classic tick timer (still not 100% calibrated, but close enough for now)
13:26:42pixelmaI usually prefer the current.txt anyway even if it means that I have to update the page manually from time to time
13:26:56TheSevenand no, it didn't fix playback
13:27:34TheSeveni did some IRQ load profiling which ended up at 5 milliseconds of time in the PCM IRQ handler over like a minute of playback
13:27:38TheSevenso that's not the cause either
13:27:47TheSevenwhere else might it get stuck?
13:27:57TheSevenshould i try to figure out tick timer IRQ load?
13:28:14Stephenc___pixelma using firefox here too, must try IE
13:28:28TheSevenit's a shame that the rockbox scheduler doesn't do some basic thread and kernel load profiling at least
13:28:38TheSeven(roughly like what embios does)
13:28:40Stephenc___TheSeven i guess it didn't fiz the Thick 160gb models either ;o) haha
13:28:54CIA-7r28988 build result: All green
13:29:34TheSevenwhat else might be causing this playback trouble?
13:29:53Stephenc___pixelma working in IE must be my old version of firefox, strange tho coz the other days it worked.
13:30:12TheSeveni'd really like to know who's eating all that CPU time...
13:32:11 Join dfkt [0] (dfkt@unaffiliated/dfkt)
13:35:40 Join TheLemonMan [0] (
13:38:03 Quit Stephenc___ (Quit: CGI:IRC)
13:38:53 Join bluebrother [0] (
13:38:53 Quit bluebrother (Changing host)
13:38:53 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
13:42:00 Quit kevku (Ping timeout: 272 seconds)
13:42:11 Quit bluebroth3r (Ping timeout: 276 seconds)
13:44:10 Join petur [0] (~petur@rockbox/developer/petur)
13:46:28linuxstbTheSeven: I'm reading - does it really destroy all your data? What exactly is happening when you follow those instructions - is iloader being written to NOR?
13:46:52linuxstbyes to which question?
13:47:14TheSevenit will be flashed to NOR and the HDD will be reformatted, as apple seems to be doing something funky that's preventing us from mounting it
13:47:32linuxstbWell, that's nice of them...
13:47:55TheSevenmight be related to them hiding away the firmware partition and thus offsetting things
13:48:13TheSeveni haven't inspected that in detail yet, i just wanted to get around it for now
13:48:37TheSevenalso it doesn't really hurt much as we don't have dualboot yet
13:48:50 Join kevku [0] (~kevku@2001:7d0:0:f000::135d)
13:49:25TheSeven(the cause for dualboot failing seems to be the OF locking up waiting for an I2C bus access to finish - and I have absolutely no clue why)
13:49:55linuxstbDo you keep the Apple code in NOR?
13:51:34 Quit parafin (Ping timeout: 260 seconds)
13:52:33TheSevenno, it will be written to norflash.bak and then wiped
13:55:27 Nick Guinness` is now known as Guinness (
13:55:34 Join parafin [0] (
13:56:01 Join Buschel [0] (
14:10:03 Quit benedikt93 (Quit: Bye ;))
14:13:33n1sTheSeven: did you try with a dummy pcm driver (that doesn't actually do the transfers to the codec)?
14:14:00n1smy guess would still be too many interrupts but from what=
14:14:50TheSevenno, i didn't try that yet.
14:14:58TheSevenhow should such a thing behave?
14:15:12TheSevenregister a tick function that calls the PCM get more callback?
14:15:24n1si don't know exactly
14:18:04TheSevenhm, might pcm_get_bytes_waiting returning very funky values be the cause?
14:20:54TheSevenit actually seems to be pcm_get_bytes_waiting that's eating the cpu cycles
14:20:59Buschelfixed it? ;)
14:22:18TheSevenno, i have absolutely no clue why it happens, but i have proof that this function is going nuts
14:23:05n1sthe codec thread seems to use that function to determine the elapsed time so i'd guess wrong values would confuse it
14:23:19Buschelwell, the implementation for S5L8702 is the only that has a while-loop inside
14:23:42TheSeven (what you just looked at)
14:23:57TheSeventhe splash shows like 300000 iterations sometimes
14:24:38TheSevenhm, that's not trivial to fix
14:24:39Buschelthis sounds like you've got it
14:25:08TheSevenyeah, i know what it's doing and why, but it won't be easy to change that
14:25:13*TheSeven heads off for lunch
14:26:21***Saving seen data "./dancer.seen"
14:27:26 Join WonTu [0] (
14:27:40 Part WonTu
14:30:52 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
14:36:00 Join sideral [0] (~sideral@unaffiliated/sideral)
14:47:24TheSevenoh yeah
14:47:35TheSeventhat function only ever terminates if the pcm buffer is empty
14:47:43TheSeventhat explains the stuttering very well
14:48:00[Saint] it an easy fix, though?
14:48:13TheSevenhowever it doesn't explain how flac can ever have worked
14:48:18Buschelwould "return remaining + (DMAC0C0LLI.control & 0xfff) * 2;" fix it?
14:48:22[Saint]I've been looking for a cheap Classic, btw ;)
14:48:37*[Saint] wants in on this port.
14:48:38TheSeventhat would fix it, but return wrong results
14:48:57TheSevenactually i need to figure out where the end of that ring buffer is :)
14:49:39Tornenowhere, it's a ring. ;)
14:51:38Buschelthe functions return value is used to check the playback latency. what would go wrong if this would results in a too small number?
14:52:03Buschelbaaah, typos, typos, typos...
14:53:21Buschelwould it really be an issue?
14:54:26 Join mgue [0] (
14:55:38CIA-7New commit by 03theseven (r28989): Fix iPod Classic playback slowness
14:55:53TheSevenplayback still isn't clean though
14:56:03TheSevensometimes wrong chunks seem to slip in
14:58:02*TheSeven starts to consider the ipod classic port usable :)
14:58:08CIA-7r28989 build result: 0 errors, 1 warnings (theseven committed)
14:58:18TheSeven(not meant in terms of our tiers, but "ready for everyday use by myself)
14:58:58TheSevennow where does that warning come from
15:00:02*TheSeven needs to pun away the volatile
15:00:38CIA-7New commit by 03theseven (r28990): Fix yellow
15:02:01 Join komputes [0] (~komputes@ubuntu/member/komputes)
15:02:02*TheSeven spots some GPIO issues
15:03:21CIA-7r28990 build result: All green
15:03:47TheSevenand that annoying "playing garbage after plugging usb in charging mode" bug is back :/
15:04:45*TheSeven wonders where those periodic glitches are coming from
15:05:44TheSevena periodic glitch every three seconds
15:05:53TheSeven(sounds like some small chunk being played twice)
15:07:48TheSevenhm, three seconds is the length of the PCM buffer
15:09:57TheSevenBuschel: it would cause the playback progessbar to jump around, and the only reason why it wouldn't cause playback corruption is the lacking support for pcm double buffering in the pcmbuf code
15:10:40BuschelTheSeven: after your changes to the us timer -> what do the mem and codec tests say?
15:11:06TheSevenyou mean the tick timer?
15:12:32TheSevenDRAM: 170.6MB/s read, 345.9MB/s write, 355.5MB/s memset, 216.9MB/s memcpy
15:12:55TheSevenIRAM: 206.4MB/s read, 290.9MB/s write, 290.9MB/s memset, 246.1MB/s memcpy
15:14:39Buschelthis is a far more than the 8700/8701 reaches...
15:16:10 Quit anewuser ()
15:16:32TheSevena factor of ~10 IIRC?
15:16:43BuschelIRAM: rd 2x, wr 4x DRAM: rd 3x, wr 9x
15:16:52Buschelabout this...
15:17:18 Join Keripo [0] (
15:17:21TheSevenmDDR seems to pay off :)
15:18:00Buscheltest codec results should also reach wuit
15:18:20Buscheltest codec results should also reach quite interesting numbers
15:19:10TheSevenhm, now why does the pcmbuf wrapping cause glitches? weird chunk sizes?
15:22:03TheSevenwow. splashf works from an IRQ handler o.0
15:23:17 Quit wodz (Quit: CGI:IRC)
15:23:25n1sheh as long as it doesn't sleep, right? also IIUC it's not guaranteed to work though, the ui thread could draw over it at any time
15:27:39 Quit Buschel (Quit: ChatZilla 0.9.86 [Firefox 3.6.13/20101203075014])
15:28:34 Quit stoffel (Ping timeout: 240 seconds)
15:31:06TheSevenyes, the UI thread starts throwing things into the top left corner sometimes, but it works well enough for debugging :)
15:34:06 Join sasquatch [0] (
15:36:55 Quit Keripo (Quit: Leaving.)
15:42:57 Quit petur (Quit: *plop*)
15:51:21 Join anewuser [0] (anewuser@unaffiliated/anewuser)
15:53:11 Join webguest67 [0] (
15:53:21webguest67Hi, I need some help
15:54:22n1swebguest67: if you tell us what your problem is it's a lot easier to help
15:55:51webguest67Ok, i got rockbox for ipod video, but when i go to use the ipod patch thing, it says there is no ipod connected
15:55:57webguest67Ok, i got rockbox for ipod video, but when i go to use the ipod patch thing, it says there is no ipod connected
15:56:04DBUGEnqueued KICK webguest67
15:56:04webguest67Ok, i got rockbox for ipod video, but when i go to use the ipod patch thing, it says there is no ipod connected
15:56:51n1swhich "ipod patch thing" ?
15:57:01webguest67ok, the ipod patch says there is no ipod connected.
15:57:15 Quit webguest67 (Client Quit)
15:57:24 Join webguest11 [0] (
15:57:44webguest11ipod patch says there is no ipod connected
15:57:53 Part part ("ugh - repeats")
15:58:13 Join Keripo [0] (
15:58:37n1swebguest11: is that "ipod patch thing" ipod patcher or rbutil? Are you *sure* it's an ipod video and not a "classic"?
15:59:05n1soh and have you installed ipod linux or done anything else funky to the partition table?
15:59:19webguest11no, this is all i want to do
16:00:03webguest11and i got the patch off the online manual....
16:00:13webguest11and i have windows
16:00:20webguest11and it is an ipod video
16:02:22webguest11I'm trying again as we speak
16:03:06LloreanWhat capacity?
16:03:39webguest118 gb
16:03:57LloreanThat's not an iPod Video.
16:04:05LloreanAre you perhaps referring to the "iPod Nano Video"?
16:04:28LloreanIt's important to know exactly which player you have before you start.
16:04:34webguest11it may be an ipod nano 3rd gen
16:04:40LloreanThe iPod Video came in 30, 60, and 80GB capacities.
16:04:45LloreanRockbox does not work on the iPod Nano 3rd gen.
16:04:48webguest11in fact, i'm sure it is now
16:04:50[Saint]There's an "iPod Nano Video"?!?
16:04:56n1s[Saint]: no
16:05:07[Saint]what naming scheme is that pulled from...?
16:05:11Llorean[Saint]: That's what some people call certain iPod Nanos, though it's an unofficial name.
16:05:13n1sther are nanos that can play videos thpo
16:05:25[Saint]n1s: Errr...wrong.
16:05:31n1sthe same as classic color etc :)
16:05:33webguest11web going funky, brb
16:05:35[Saint]dammit, sorry.
16:05:36 Quit webguest11 (Quit: CGI:IRC)
16:05:44 Join webguest17 [0] (
16:05:46[Saint]I read an imaginary "no" in there :P
16:06:12webguest17so, is there no rockbox for ipod nano 3rd gen?
16:06:47 Quit Keripo (Quit: Leaving.)
16:07:06webguest17ok, so this is a no go
16:07:58webguest17so, if i were to ty the 2nd gen, could it possibly work?
16:08:11[Saint]Try some reading.
16:08:22[Saint] lists our supported devices.
16:08:30 Quit webguest17 (Client Quit)
16:10:14 Quit sasquatch (Ping timeout: 240 seconds)
16:11:18 Join sasquatch [0] (
16:16:19CIA-7New commit by 03theseven (r28991): iPod Classic: Fix remaining PCM issues and add DMA information to the HW debug screen
16:16:31TheSevenfinally glitch-free playback...
16:16:35 Join sideral1 [0] (~sideral@unaffiliated/sideral)
16:18:49CIA-7r28991 build result: All green
16:19:28 Quit sideral (Ping timeout: 240 seconds)
16:26:25***Saving seen data "./dancer.seen"
16:26:26 Quit sideral1 (Remote host closed the connection)
16:27:10 Join sideral [0] (~sideral@unaffiliated/sideral)
16:30:33CIA-7New commit by 03theseven (r28992): Add LCD type to iPod Classic HW debug screen
16:33:14CIA-7r28992 build result: All green
16:36:41 Quit sideral (Remote host closed the connection)
16:37:36 Join sideral [0] (~sideral@unaffiliated/sideral)
16:46:58 Quit sideral (Ping timeout: 240 seconds)
16:47:42 Join sideral [0] (~sideral@unaffiliated/sideral)
16:47:52 Quit guymann (Quit: new kernel)
16:48:43 Quit MethoS- (Remote host closed the connection)
16:50:48 Join Horscht [0] (~Horscht@xbmc/user/horscht)
16:57:11 Quit sideral (Remote host closed the connection)
16:58:12 Join sideral [0] (~sideral@unaffiliated/sideral)
16:59:17rasherNo Classic on the front page?
16:59:40 Join guymann [0] (
17:00:08[Saint]rasher: IIRC the change was pushed to SVN, but the page needs updating.
17:00:54 Join MethoS- [0] (~clemens@
17:01:37 Quit MethoS- (Remote host closed the connection)
17:03:01 Join kugel [0] (~kugel@
17:03:02 Quit kugel (Changing host)
17:03:02 Join kugel [0] (~kugel@rockbox/developer/kugel)
17:03:22 Join MethoS- [0] (~clemens@
17:04:10 Quit MethoS- (Remote host closed the connection)
17:06:36 Join factor [0] (~factor@
17:07:03 Join MethoS- [0] (~clemens@
17:07:19 Quit MethoS- (Remote host closed the connection)
17:07:26 Quit sideral (Remote host closed the connection)
17:08:17 Join sideral [0] (~sideral@unaffiliated/sideral)
17:08:22 Quit kugel (Read error: Connection reset by peer)
17:09:39 Join panni_ [0] (
17:09:46 Join kugel [0] (~kugel@
17:09:48 Quit kugel (Changing host)
17:09:48 Join kugel [0] (~kugel@rockbox/developer/kugel)
17:12:56 Join MethoS- [0] (~clemens@
17:13:47 Quit MethoS- (Remote host closed the connection)
17:17:29 Join MethoS- [0] (~clemens@
17:17:50 Quit MethoS- (Remote host closed the connection)
17:21:20 Join Judas_PhD [0] (
17:21:48 Join MethoS- [0] (~clemens@
17:24:27 Quit sideral (Ping timeout: 246 seconds)
17:26:31 Join stoffel [0] (
17:27:54 Join sideral [0] (~sideral@unaffiliated/sideral)
17:32:07 Join captainkewll [0] (2669ecc2@gateway/web/freenode/ip.
17:33:05 Join bertrik [0] (
17:33:05 Quit bertrik (Changing host)
17:33:05 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
17:33:29 Join {phoenix} [0] (
17:37:54 Join {-phoenix-} [0] (
17:37:56 Quit sideral (Remote host closed the connection)
17:38:43 Join sideral [0] (~sideral@unaffiliated/sideral)
17:39:38 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
17:40:55 Quit guymann (Ping timeout: 264 seconds)
17:41:40 Quit FOAD (Quit: I'll be back)
17:42:07 Quit {phoenix} (Ping timeout: 264 seconds)
17:44:55 Quit sideral (Ping timeout: 240 seconds)
17:46:00 Join FOAD [0] (~dok@
17:52:39 Quit elcan (Read error: Connection reset by peer)
18:02:39 Quit TheLemonMan (Quit: free(me))
18:04:28 Quit mgue (Quit: Lost terminal)
18:09:36 Quit tails__ (Ping timeout: 246 seconds)
18:14:58 Join krabador [0] (
18:15:51krabadorhi people, by sometimes i'm trying rockbox on android, i'm trying the svn version available on rockbox site
18:16:31krabadormy lead use is music playback
18:16:44krabadorfor flac files is the best way at the moment
18:17:33krabadorby the way when i'm using a file browser, i can't chose rockbox on "open with " option
18:17:49 Join elcan [0] (
18:17:57krabadorisn't yet implemented, or it's a problem of mine?
18:19:42 Join Guinness` [0] (
18:20:14Tornerockbox isn't very integrated into android as an app yet
18:20:23Torneso that kind of functionality is not implemented
18:20:35 Quit Guinness` (Read error: Connection reset by peer)
18:22:55 Quit Guinness (Ping timeout: 240 seconds)
18:24:31 Join Guinness [0] (
18:26:26***Saving seen data "./dancer.seen"
18:28:50 Quit B4gder (Ping timeout: 240 seconds)
18:30:58 Join cormon [0] (
18:31:48cormonIs .m4a supported in rockbox
18:32:17 Join B4gder [0] (~daniel@rockbox/developer/bagder)
18:32:27n1scormon: yes
18:32:43n1s(as a container for aac or alac IIRC)
18:32:58cormonit does not seem to work on my cowon D2.
18:33:53Torneare the files DRM'ed/protected/encrypted?
18:34:22LloreanRockbox also can't play very long ones.
18:34:27LloreanAnd depends o
18:34:41Lloreandepends on the metadata being organized properly (at the start of the file?)
18:35:42 Join LambdaCalculus37 [0] (~3f74f70d@rockbox/staff/LambdaCalculus37)
18:36:05cormonthe files are not DRMed. They are just over two hours long.
18:36:23LambdaCalculus37B4gder: Ping
18:36:45 Join FlynDice [0] (
18:36:54cormonthat could be my problem.
18:38:01cormonis there any solution to playing long m4a files on rockbox or will i have to split the file?
18:39:30LloreanSplit, or convert to another format.
18:39:39LloreanI don't believe any other supported format has a length restriction.
18:39:58n1sif you have foobar2k there's some feature to "optimize" the layout of the files that might help
18:40:26LloreanYes, that will solve the problem if the issue is the metadata location.
18:41:20 Join sbmull [0] (
18:42:42sbmullI have written a LUA Flash Card program for my Sansa Clip+. Please see FS #11864 - Working LUA Flashcard Program
18:44:46sbmullIt uses export files from the popular ANKI ( I have posted code, a card file and screenshots.
18:45:08sbmullI use it for studying Russian. Unicode works great!
18:45:23 Quit FlynDice (Remote host closed the connection)
18:45:41 Quit sbmull (Client Quit)
18:46:13 Join sbmull [0] (
18:47:48amee2kdoes the Application target compile again?
18:48:03gevaertssbmull: if you want sqlite, it may be possible to integrate FS #8852 in lua
18:48:09amee2ki don't remember who, but someone was making modifications to it yesterday that resulted in compile errors
18:49:39gevaertsamee2k: no fix has been committed yet. Reverting the two commits that broke this is still the best you can do
18:49:57 Quit shai (Quit: Leaving)
18:50:32LambdaCalculus37TheSeven: Ping, can you send me a build for the iPod Classic built on the most recent SVN?
18:51:07CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
18:51:07*TheSeven just updated the one on, if you mean an installer build
18:51:21LambdaCalculus37TheSeven: Just a Rockbox build.
18:51:32TheSeventhat has already been added to the build system :)
18:52:20LambdaCalculus37TheSeven: Thanks. :)
18:53:58AlexP[Saint]: Wassup?
18:54:59[Saint]AlexP: Sorry, nevermind...I thought I found a tag missing in the manual but it was just in a place I didn't expect it to be at the time.
18:55:13[Saint]and, unfortunately you can't "un-ping" someone.
18:55:24AlexPheh, no worries :)
18:55:42pixelmawhich tag was it?
18:56:34 Join guymann [0] (
18:57:01LambdaCalculus37TheSeven: Help my bleeding ears here... you said that you had gotten volume control working on the Classic before, right?
18:57:43 Quit kugel (Ping timeout: 264 seconds)
18:58:22[Saint]pixelma: %cc "has RTC?"
18:58:48[Saint]I was looking in Time & Date, but I forgot about the Hardware Capabilities section.
18:58:52pixelmaisn't it in the hardware present section?
18:59:25LambdaCalculus37TheSeven: Adjusting the sound controls affects the output volume. I'm trying to raise up the bass on my Classic in the sound setting menu and it nearly blew my brains out.
18:59:26TheSevenLambdaCalculus37: it was a matter of adjusting some rather insane ifdef
19:00:06TheSevenhm, interesting
19:00:24TheSevenapparently enabling the DSP increases the volume
19:00:59TheSevenhm, or it's just some reg mixup
19:01:01pixelma[Saint]: Time & Date was a bit silly for the targets that do not have an RTC - this section was only enabled for this one tag and I guess that was confusing too
19:01:01TheSevenlet me check
19:01:04LambdaCalculus37TheSeven: Same behavior applies when I try adjusting the treble settings.
19:01:05 Quit cormon (Quit: ChatZilla 0.9.86 [Firefox 3.6.13/20101203075014])
19:01:17[Saint]How would you get around that? Pre-cut when a DSP is enabled?
19:01:44TheSeveni'm pretty sure it's just a bug somewhere
19:01:46LambdaCalculus37[Saint]: IIRC that's the way to do it.
19:02:11 Quit pjm0616 (Read error: Connection reset by peer)
19:04:17[Saint]pixelma: Yes, it was just my train of thought - "%cc - is there an RTC? - clock - time - time & date - why isn't it there? - it's not there!"
19:04:23LambdaCalculus37TheSeven: But on the bright side, great work on speeding up audio decoding. :)
19:04:31[Saint]and then I found the Hardware Capabilities section :P
19:06:41 Join saratoga [0] (9803c6dd@gateway/web/freenode/ip.
19:07:06TheSevenLambdaCalculus37: it's still a mystery how FLAC could work before that fix
19:07:19saratogaLlorean, n1s: (catching up on logs) IIRC the foobar optimize function shouldn't be needed anymore for M4A files
19:07:41saratogaalthough theres basically no chance of a 2 hour AAC file playing anyway
19:07:43TheSeventhat bug made that "get remaining bytes" function loop until the pcm buffer was almost empty
19:08:01saratogaah, makes sense why testcodec worked fine then!
19:08:17saratogai'm curious though, i saw you fixed the timer, does that mean your test_codec results were inaccurate?
19:08:34TheSevenif they depend on the tick timer, probably yes
19:08:43Lloreansaratoga: Is there some reason we can handle arbitrary lengths of most formats, but not AAC?
19:08:47TheSevendivide the MHz numbers by ~1.6 in that case :)
19:08:53 Quit Topy (Read error: Connection reset by peer)
19:08:54LloreanIs it a format limitation, or just something "not finished yet" in Rockbox?
19:08:58saratogai expect them to be wrong, but i'm not sure about the timmer
19:09:13*TheSeven might do a full codec test tonight
19:09:26saratogaLlorean: yeah because the mp4 spec is a mess and the spec isn't free either, so i'm not sure what the "right" fix is to long files
19:09:33TheSevenit would be nice if test_codec would use the usec timer for that
19:09:54saratogai can hack it up so that long files play (at least from apple) but i have no idea if the parser is even correct at that point
19:10:34saratogaright now we use a look up table to keep track of where individual mp4 frames are in the bit stream, but if the file is more then 45 minutes long you run out of ram to store the lookup table
19:11:19saratogaTheSeven: if those memory results are right then I expect you to be quite a bit faster then AMSv2, which could do flac in under 10MHz
19:12:32sbmullgevaerts: looks like FS #8852 has been closed.
19:12:34Lloreansaratoga: Might it make sense to take your "hacked up"version and put it in the 'official test builds' section and see if anyone can find files it doesn't like?
19:12:38 Quit bmbl (Quit: Verlassend)
19:12:45*TheSeven would really like to know how fast the nano4g would be
19:12:55gevaertssbmull: yes, because we don't plan to commit that without any code that uses it
19:13:14TheSeventhese little things might well outperform the beast
19:13:17saratogayes, thats on my todo list, just pretty far back since no one uses aac all that much
19:14:18LloreanAAC-HE seems to be pretty popular for audiobooks performing pretty well at low bitrates.
19:14:31*LambdaCalculus37 sees that the nano3G and the Classic1G are not quite as similar hardware-wise as he was initially thinking
19:15:11TheSevenLambdaCalculus37: different PMU, different codec, different RAM
19:15:20TheSevenapart from that, they should be fairly similar
19:15:24LambdaCalculus37TheSeven: Yup.
19:15:40saratogayeah, i'm just really hesitant to create yet another 'almost but not quite right' mpeg implementation
19:17:02 Join DerPapst1 [0] (
19:17:04 Join towolf [0] (
19:17:22LambdaCalculus37TheSeven: And I also know that emBIOS doesn't work on the nano3G as well.
19:17:36TheSevenwhich is due to the PMU and RAM
19:17:57TheSevenand some differences in the GPIO setup
19:18:04 Join LambdaCalculus38 [0] (
19:18:28GodEaterdon't know if anyone else found this or not yet, but my rockbox article in Linux Journal is now publically available :
19:18:38 Quit LambdaCalculus38 (Client Quit)
19:18:43GodEaterincreasingly inaccurate as time passes of course ;)
19:18:55 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
19:19:11gevaertsGodEater: this means you should update the wiki :)
19:19:11saratogaGodEater: nice
19:19:19GodEaterwhich page gevaerts ?
19:19:22LambdaCalculus37GodEater: But still an excellent article all the same. :)
19:19:29towolfi might be missing something, recording from mic on sansa clipv2. i cranked the gain up to the maximum, and the audio clips like crazy, but the sample values in the written wave only go up to 0.25 of the full range.
19:19:30LambdaCalculus37GodEater: ArticlesAboutRockbox
19:19:48GodEaterI'll go do that
19:19:50saratogai was thinking something like that for the IEEE student magazine might be nice, they've done them in the past on open source firmware for cameras, and IEEE readers are quite likely to be interested in firmware hacking verses normal people
19:20:14saratogatowolf: 14 bit ADC would be my guess
19:21:01towolfsaratoga, so 2 bits of headroom are lost? is this normal?
19:21:23saratogatowolf: you mean compared to a 16 bit ADC?
19:21:48towolf... that these little devices don't record at full level
19:21:55LambdaCalculus37TheSeven: The only other DAP I saw in this store was a nano5G, but I know that's totally out of the loop right now. Not a shred of custom code for it.
19:21:56towolfs/level/bit depth
19:23:18saratogatowolf: we should probably automatically increase the volume of the output file by 2 bits, but otherwise yes thats normal
19:24:00*GodEater updates said page.
19:24:26towolfsaratoga, ok, thank you. i was wondering about this. who knew that lower bit depth would manifest in lower volume. but it makes sense. once normalized it will have reduces resolution.
19:25:02*LambdaCalculus37 has LCD type 1 on his Classic according to the View HW Info screen
19:25:04 Join Topy44 [0] (
19:25:32TheSevenLambdaCalculus37: so it works right?
19:25:42 Join Buschel [0] (
19:25:42LambdaCalculus37TheSeven: Like a charm.
19:26:00 Quit MethoS- (Remote host closed the connection)
19:26:21LambdaCalculus37TheSeven: CPU freq is always at 216MHz and is stuck at 100% boost at all times.
19:26:37TheSevenyes, there is no clock scaling driver yet
19:29:37Buschelsaratoga: you saw my comment in FS #11858?
19:29:43 Quit LambdaCalculus37 (Quit: back to work)
19:29:51sbmullgevaerts: What would be involved to make the sql api open to lua
19:30:22gevaertssbmull: no idea. I've never done any lua work...
19:31:02n1slinuxstb: turns out that the actual /lib/cpp symlink isn't needed and the g++ package doesn't create it on cygwin it was just that the error message was a bit unhelpful
19:31:44n1sg++ (or some other c++) compiler seems to be required though
19:32:35sbmullgevaerts: This was my first lua work.
19:33:41sbmullgevaerts: I was suprised how easy it was. Faster to test and bug than c too.
19:34:51TheSevenhm, one issue with the tone control is probably the corner frequencies for bass/treble
19:35:02TheSevenbass is at 50Hz, treble at 5kHz, which is probably both too low
19:35:43 Join Rob2222 [0] (
19:36:53 Join kugel [0] (~kugel@rockbox/developer/kugel)
19:37:52 Quit towolf (Quit: Ex-Chat)
19:38:45 Quit sbmull (Quit: CGI:IRC)
19:42:43B4gderfor the record, if you want to get in touch with Zagor and he's not here, he's generally available by email
19:45:40CIA-7New commit by 03gevaerts (r28993): Add some app_*() wrappers for file IO functions to make app_ work the same as sim_, thereby fixing application builds
19:45:55gevaertskugel: can you check if I didn't break android (more)?
19:47:27 Quit AlexP (Remote host closed the connection)
19:47:44GodEateryou broke android?
19:47:54GodEaterlike, the whole thing?
19:47:57 Quit krabador (Ping timeout: 260 seconds)
19:47:58*GodEater is impressed
19:48:09CIA-7r28993 build result: All green
19:48:28GodEaterI know you're not a huge fan of it - but that seems an extreme reaction
19:49:19gevaertsI just want to make sure I *didn't* break it!
19:49:38GodEateryou said "(more)" - which I took to me, I broke it a bit already
19:50:39*GodEater fears that with the ambiguousness of this statement, he may have just opened a semantic can of worms with gevaerts
19:50:59*TheSeven shoves that discussion to #rockbox-community
19:51:04 Join AlexP [0] (~alex@rockbox/staff/AlexP)
19:51:40GodEaterthanks TheSeven ;)
19:53:37kugelcan we use aliases (instead of wrappers) for library functions?
19:54:28gevaertshm, yes
19:55:39Buschelsoap: you experienced the playback/buffer issue again?
19:56:18Buschelsoap: ...or even found a way to reproduce it?
19:58:40gevaertskugel: the only problem is that I keep getting confused by function pointer syntax
20:00:58gevaertskugel: feel free to have a go
20:01:25 Join krabador [0] (
20:01:57saratogaBuschel: i did but was a little confused, has anything even changed between those revisions?
20:02:53gevaertsamee2k: you should be able to build again
20:03:25Buschelsaratoga: yes, r28937 changed some error handling in apps/metadat/rm.c
20:04:15 Quit factor (Read error: Connection reset by peer)
20:04:36Buschelsaratoga: rolling back this change does not seem to be the best solution as this change was made to fix a ticket (by you) regarding a crash in the database parser
20:06:16 Join factor [0] (~factor@
20:06:24[Saint]gevaerts: Does the claim of App-fixedness extend to CygWin?
20:06:47gevaerts[Saint]: there shouldn't have been anything cygwin-specific about this
20:08:15Buschel[Saint]: sorry to bother you again -> was the tearing on you iPod color LCD introduced by my latest changes? or was it the same before?
20:14:34 Join towolf [0] (
20:14:36 Quit AlexP (Remote host closed the connection)
20:15:00[Saint]I will find the time to go through the revisions and check which (if any) patch introduces the tearing this afternoon.
20:15:07 Quit stoffel (Ping timeout: 264 seconds)
20:15:27[Saint]I have a fairly busy day, but I can get this done for you...I know you're wanting to commit these things.
20:15:31[Saint]Buschel: ^
20:16:09Buschel[Saint]: it has been committed already. you've mentioned this tearing some days ago (when CPU is boosted)
20:16:10towolfsaratoga, regarding the 14-bit to 16-bit issue. do i understand right that i just need to bit-shift by 2? are you aware of any tools that do this?
20:16:59[Saint]Buschel: Um...what has been committed already?
20:17:05[Saint]and, when?
20:17:37Buschel[Saint]: all available LCD optimizations (except some YUV stuff which is not relevant for the tearing you described)
20:17:53towolfi further assume that a standard normalization would achieve almost the same thing but would introduce minor rounding errors? so a bit-shift would be simpler and cleaner?
20:18:55*[Saint] wonders why he didn't get any conflicts when he updated his SVN checkout.
20:19:11Buschel[Saint]: maybe because the changes were identical
20:19:23[Saint]Ah, possibly.
20:19:30gevaertstowolf: is the quality of that microphone really good enough for these possible rounding errors to matter?
20:20:54towolfgevaerts, not really, but in normalization it maximizes according to the single maximum peak sample
20:21:42towolfthat is probably the main issue. i just need to multiply the sample values by 4. as far as i understood things correctly. right?
20:22:07saratogaBuschel: i don't see how that change breaks anything though?
20:22:30 Quit t0rc (Remote host closed the connection)
20:22:47towolfany linux tools that do this?
20:23:24*TheSeven wonders what the hell is going on with tone control
20:25:14Buschelsaratoga: I have build a simulation with rm.c at r28936 and at r28937. it made the difference
20:25:42Buschelsaratoga: you have seen mt lately?
20:26:29***Saving seen data "./dancer.seen"
20:27:15saratogai'll IM him
20:28:14TheSevenwhich corner frequencies would you recommend?
20:28:20TheSevenbass: 50/100/200/250Hz
20:28:33TheSeventreble: 5/6/10/15kHz
20:28:43TheSevener, 5/7/10/15
20:28:47saratogaBuschel: i don't have the sim handy right now, but if you want you could try changin gthe line to "if(temp < 0 || skipped < 0)" instead of "if(temp < 0)"
20:28:49 Join wodz [0] (
20:29:01saratogathats the only thing i see that might make a difference
20:29:18TheSevenactually i'd like this to be configurable but I don't think we have settingss for that right now
20:29:41TheSevenwhat order of magnitude of values are we using on other targets?
20:30:08pixelma[Saint], gevaerts: the app build on cygwin or mingw had a different problem before, so I guess it's back to that one now (haven't tried yet though)
20:30:23wodzWhat program would you recommend to inspect metadata in mp3? And BTW. how is replaygain handled from the point of mp3 file? Should the file have some special tags?
20:30:43saratogawodz: it has an id3v2 tag for replaygain
20:30:59saratogato look at metadata i usually just use foobar2000 on windows or ffmpeg on linux
20:31:11[Saint]pixelma: I'm building now...fingers crossed.
20:31:24wodzsaratoga, thx
20:31:29[Saint]I'll have to leave before it finished though, so results will be delayed somewhat.
20:31:49pixelma[Saint]: the error appeared when building the exe, so near the end
20:32:01Buschelsaratoga: will do in a few minutes
20:32:02[Saint]yes, I remember this also.
20:32:27[Saint]pixelma: I guess I'll just hope that it magically fixes itself.
20:32:33n1sTheSeven: why not make it configurable?
20:32:44saratogaBuschel: actually i can test on device, i'll look and see if anything is obviously wrong
20:32:53pixelmaI'll try cross-compiling in my VM now
20:33:00 Quit towolf (Quit: Ex-Chat)
20:33:03TheSevenn1s: do we have the framework for that?
20:33:12n1sTheSeven: oh, missed you said that, the beast has a hw eq
20:33:12TheSeveni don't really want to dig into the settings engine right now
20:33:52n1sthe beast hw eq menu is in menus/audiohw_eq_menu.c
20:33:58wodzsaratoga: could you give me some hint? ffmpeg man page is rather big and video oriented
20:34:48n1shmm, not a great fit if i understand the capablities of your codec
20:35:48TheSeveni can basically set the corner frequency and gain of a bass and a treble shelving filter
20:35:48n1sTheSeven: are those frequencies cutoff freqs for bass and treble filters?
20:35:58n1sah, ok
20:36:22n1sso just popping in 2 extra settings for that would be simpler i think
20:36:23TheSeventhe gain is handled in a sensible way already, just not the corner frequency
20:36:46TheSeventhe problem with this is that those settings have non-linearly distributed target-dependent values
20:37:13TheSevenwhat would you recommend as default values for them?
20:37:15saratogawodz: i think it prints the metadata if you do ffmpeg -i file.mp3 out.wav
20:37:31saratogathats all i ever do with it, if you need something more then that i'm not sure
20:37:33TheSeveni would just hardwire that for now until someone thinks about a way to add a setting for it
20:38:53 Nick Ypsy is now known as YPSY (
20:39:19wodzsaratoga, <- I don't see any meta data in this output
20:39:36n1sTheSeven: the sw eq uses 80Hz and 5.3kHz as defaults so something close to them maybe
20:40:10 Join sideral [0] (~sideral@unaffiliated/sideral)
20:43:09saratogawodz: for me it prints below "Stream #0.0: Audio:"
20:43:18saratogamaybe the file has no valid metadata?
20:43:57wodzit has
20:44:04saratogathen i'm not sure
20:44:22saratogaBuschel: trying those tracks they don't even hit the code path
20:45:13Buschelsaratoga: seems we're somehow not on the same page...
20:45:30saratogai used cook 64 mono and stereo
20:45:34saratogais that the right files?
20:46:01Buschelsaratoga: commenting line 382 (="return -1;") does the trick.
20:46:36Buscheldo you use the same revision of rm.c?
20:48:24saratogano its an older check out, but its the one i used to commit the rm.c changes int he first place, so it should be fine i think
20:48:26saratogai can try updating
20:48:51saratogahmm actually i probably shouldn't since this check out has all my libmad work in it
20:49:39Buschelsaratoga: mt submitted a fix after yours. the patch to fix the svn issue is this ->
20:49:39saratogawhat i don't understand is that the only way that -1 can ever be hit is if the file doesn't have a recognized FOURCC or if the extradata length is negative
20:49:47Buschel(which is more or less reverting mt's change)
20:49:56 Quit factor (Ping timeout: 264 seconds)
20:50:29saratogawhat is the "Unknown header signature " it prints?
20:50:53saratogaso the file has no fourcc
20:51:07saratogai would assume that means its corrupt, but evidently not if we used to play it
20:51:57 Join thegeek_ [0] (
20:52:10 Join afk [0] (~Dre@
20:52:31 Quit scorche (Disconnected by services)
20:52:41 Join scorche` [0] (~scorche@rockbox/administrator/scorche)
20:52:59 Join AlexP [0] (~alex@rockbox/staff/AlexP)
20:53:42 Join factor [0] (~factor@
20:55:15 Quit Dreamxtreme (Ping timeout: 255 seconds)
20:55:43 Quit thegeek (Ping timeout: 255 seconds)
21:00:10Buschelhmm, in fact the "" comes from the (buggy) fourcc2str-conversion.
21:00:33 Join TheLemonMan [0] (
21:01:07 Nick AlexP is now known as AlexP_ (~alex@rockbox/staff/AlexP)
21:01:10 Nick AlexP_ is now known as AlexP (~alex@rockbox/staff/AlexP)
21:02:39pixelmaso, the "make" for a cross-compiled app build seems to work now and I thought that I need to do a "make install" then, or do I miss something?
21:03:39pixelma"make install" failed for me because it uses a wrong dir it seems
21:05:02*Buschel 's iPod Video again hung up during playback using plain r28992
21:06:21[Saint]pixelma: "make && make install" should be all that is needed IIUC
21:08:51 Join skx [0] (~skx@unaffiliated/skx)
21:09:18[Saint]pixelma: What OS are you building in atm?
21:09:26pixelmamake install wants to install the build in my "/usr/local" dir which it isn't allowed to and I think it won't find the things needed
21:09:49pixelma[Saint]: linux VM but I chose to cross-compile for Windows
21:10:32[Saint]pixelma: I'm getting failuers with speex on CygWin...if you can, it would be nice if you could check out building voice with speex on CygWin.
21:10:42[Saint]No rush, but, I would appreciate it.
21:11:45pixelmaI already told you that it works for me, despite a warning that it couldn't find some package (would need to look it up which exactly it was but you could find it in the logs here)
21:13:01[Saint]I'm getting a bunch of: librbspeex.a(rbspeex.o):rbspeex.c:(.text+0x453): undefined reference to `_speex_nb_mode' etc.
21:13:04 Part jhMikeS
21:13:04 Join jhMikeS [0] (~jethead71@rockbox/developer/jhMikeS)
21:13:21 Join LambdaCalculus37 [0] (~3f74f70d@rockbox/staff/LambdaCalculus37)
21:13:34[Saint]No idea why, it's quite annoying. But at least I know it's a problem on my end.
21:13:39LambdaCalculus37B4gder: Can you update the front page?
21:13:42pixelmaI successfully built an M5 voice
21:13:46 Quit jhMikeS ()
21:13:54 Join jhMikeS [0] (~jethead71@rockbox/developer/jhMikeS)
21:14:59 Quit LambdaCalculus37 (Client Quit)
21:16:46 Join kadoban [0] (
21:17:34kugelpixelma: I think I've mentioned it before but the sdl app most likely won't work on windows without further work
21:18:33skxI'm using the Rockbox Utility on Ubuntu 10.10 and the autodetect doesn't work, I have provided it both with the mountpoint (/media/SANSA CLIP) and the device (/dev/sdb) and it doesn't detect my player, shoudl I be worried?
21:18:44skxhow reliable is autodetect usually?
21:19:07pixelmakugel: I remember you saying "no-one tried", so I wanted to
21:19:25[Saint]that's all I remember also
21:19:39pixelmastill curious what the cross-compile want to do with the "usr/local" dir
21:19:43[Saint]I wouldn't have bothered had I known there was no point...
21:20:48kugelit might work under cygwin
21:20:51sideralskx: I remember having seen the same behavior for my ClipV2. Just manually select the right Clip version.
21:20:59skxsideral, ok, thanks
21:21:13sideralPerhaps the autodetect code cannot discern between a Clip and a ClipV2 and so defaults to detect nothing?
21:21:33sideral(just a wild guess, I haven't looked at the code)
21:21:36skxit's v02 for sure (according to the system info on the player)
21:21:39CIA-7New commit by 03alle (r28994): pitch_detector: the menu function already returns a bool, so use the return value directly
21:22:10Buschelsaratoga: hmm, in fact the parser detects "FOURCC('M','D','P','R')" 2 times in the stream. the first time a correct ".ra" is recognized, the second time not -> abort
21:22:40saratogathat makes more sense then
21:22:44saratogai guess we need to ask mt what to do
21:23:37CIA-7New commit by 03alle (r28995): pitch_detector: remove a double 'break'
21:23:45 Join factor_ [0] (~factor@
21:23:52CIA-7r28994 build result: All green
21:24:15bertrikThe Clipv1 and the Clipv2 have different USB ids, so it should be possible to distinguish them by that
21:24:19 Quit factor (Read error: Connection reset by peer)
21:24:21pixelmasideral, skx: the Rockbox Utility uses the USB PIDs and VIDs as listed on - the page doesn't list ones for the Clipv2 so maybe this is not prepared for the Utility yet?
21:24:48skxaccording to the manual for Clip v2 automatic installation should work
21:24:57sideralThat would explain it. But I did install Rockbox with the utility
21:25:03skxit's the recommended path
21:25:40bertrikI'll check in the code
21:26:40CIA-7r28995 build result: All green
21:28:14 Join perfectdrug [0] (
21:29:48bertrikthe usb ids for both the clipv1 and the clipv2 have been in the code since SVN r25324 (March 24, 2010)
21:31:02 Quit Buschel (Read error: Connection reset by peer)
21:32:38 Join Buschel [0] (
21:33:58skxwow, that was a breeze, I mean... it just worked
21:34:39[Saint]what did you expect? ;)
21:34:53skx[Saint], 2 hours of configuration
21:34:55skxtbh L:)
21:36:23skxit's excellent, thanks for this and for the rockbox utility
21:36:28 Part skx
21:40:38 Quit saratoga (Quit: Page closed)
21:40:39CIA-7New commit by 03jethead71 (r28996): Gigabeat S/i.MX31: Take care of an interrupt priority inversion that can happen during PCM callback lockout when DVFS switches frequecies during the ...
21:42:57CIA-7r28996 build result: All green
21:47:15 Join skx [0] (~skx@unaffiliated/skx)
21:51:19 Part skx
21:53:38CIA-7New commit by 03alle (r28997): Rockbox Utility: update the Russian translation (FS #11861 by Simon Zhukovsky with minor tweaks by me)
21:53:44TheSevenwhat the hell.
21:54:29TheSeveneven if i make the audiohw_set_bass/treble functions do nothing, the volume increases greatly when increasing the bass/treble value in the sound settings menu above 0dB
21:55:31Buschelsaratoga: updated FS #11858
21:55:32jhMikeSsounds logical
21:55:44TheSevenwhy on earth?
21:55:59CIA-7r28997 build result: All green
21:56:00[Saint]One word: "Apple"
21:56:39kugelTheSeven: is the software bass/treble running?
21:56:45TheSevenno idea
21:57:05TheSeveneven if it would be, that kind of thing certainly shouldn't happen
21:57:45kugelnear 0db it works by making all other frequencies quiter, to avoid clipping
21:57:51jhMikeSTheSeven: I'd expect it to get quieter or not change at all
21:57:57Buschelcan r28877 lead to lockups on dual core engines?
21:58:24kugel(if you raise bass/treble, that is)
21:59:12TheSevenoh, I think I know what might be going on
21:59:33TheSeveni have 1.5dB gain steps for that
21:59:35jhMikeSBuschel: I would be very surprised it it would
21:59:59TheSevenand if i go to +1.5dB, it sounds like the volume went up by 15dB
22:00:09TheSevenis there something ignoring num_decimals?
22:00:25TheSevenand thus over-compensating by a factor of 9?
22:00:25 Join leavittx [0] (
22:01:40jhMikeSyou need the unscaled current_bass, current_treble
22:02:11jhMikeSif you're modelling this after gigabeat F's way
22:02:18TheSevenhow does this work with hardware bass/treble?
22:02:35TheSevenare things attenuated in software to prevent clipping in the codec's DSP?
22:02:54jhMikeSit's in tenth-db's already, so the anti-clipping attenuation doesn't need the "*10"
22:03:05jhMikeSTheSeven: indeed
22:03:18jhMikeSsort of
22:03:45jhMikeSif the hardware can prescale the volume somehow, the hw does it. if it's SW tone controls, the DSP does it
22:03:52TheSevenhm, i should probably implement hardware pre-cut then
22:04:47jhMikeSit scales the volume down by the band with the highest posive gain, or else leaves it alone if none have + gain
22:05:05BuscheljhMikeS: any idea what could cause FS #11863 ? this definately came in before the ATA changes
22:06:13TheSevenjhMikeS: so it basically scales it down in pre-cut, then applies bass/treble, and then scales it up again in the amp
22:06:36TheSeven(and scales it up by an order of magnitude too much on the classic)
22:06:46TheSevenis that upscaling also effective for lineout?
22:07:28 Join Keripo [0] (
22:08:09jhMikeSTheSeven: it should be effective but I suppose that depends on the codec itself and its setup
22:08:42TheSeveni've seen a bunch of hardwared 0dB's for lineout in sound.c, which will probably interfere with that
22:08:44jhMikeSBuschel: I haven't seen any issues myself on my targets.
22:09:37jhMikeSTheSeven: on the targets I implemented, I used DAC volume as a precut which doesn't affect any analog stuff. I don't know if you can do that there.
22:10:08TheSeveni can actually do that in two places
22:10:25 Join max131 [0] (
22:10:27jhMikeSBuschel: I'll think about it but nothing obvious strikes me
22:10:37TheSeventhere's a prescaler in the dsp, and another one after the mixer, before the tone control
22:10:54TheSeven page 26
22:11:13BuscheljhMikeS: that would be helpful. maybe you get an idea...
22:11:45jhMikeSBuschel: They mention mpegplayer which shouldn't be affected by priority changes by the pcm buffer that don't happen when it's running
22:11:51TheSeveni'll probably use MSTxVOL for that, so that one can apply a pre-cut even if the DSP is powered off
22:13:23BuscheljhMikeS: also test_codec did not load today when I experienced this. there is no clear image about what is happening in detail...
22:15:02max131so i suffered from an fs-corruption again during past day, and i'm running r28979 now
22:15:09BuscheljhMikeS: for now it did only happen for soap (iPod nano 1g) and me (iPod Video). I did not happen on my iPod nano 2g yet. that's why I am guessing it might be connected to dualcore. or maybe some ATA DMA stuff
22:15:34max131that's on a fuze-v1, anybody else seeing this issue occasionally?
22:16:26B4gderfront page updated
22:16:47jhMikeSTheSeven: PMIXxVOL?
22:18:27max131i'd like to help and pinpoint where the problem is, but i'm not a coder, and i would like to get the fuze more stable
22:18:33TheSeventhat's only available if the DSP is powered on and has a smaller range
22:18:46TheSevenand from a clipping point of view MSTxVOL is early enough
22:19:19TheSevenand, as the tone controls are effective for all signal sources, i'd say the precut should be as well
22:19:59max131i thought the whole issue started with r28834, but apparently the fs-corruption re-occured this morning, after having the storage formatted using the original firmware
22:20:35jhMikeSBuschel: I've played my H10, e200 and 3g without issues
22:21:37BuscheljhMikeS: it happens randomly. just let the devices play and leave them alone for a while
22:22:24jhMikeShmmm...I was listening to e200 last night for several hours...I'll try though. Which one do you prefer of those I test?
22:22:53Buschelinteresting. which build did you use?
22:25:20max131so i guess there still must be a problem in the USB or block layer in rockbox, or what else could cause a 'delayed write failed'?
22:26:33***Saving seen data "./dancer.seen"
22:26:38BuscheljhMikeS: btw, I would prefer if you take the H10. the issue might be connected to ATA DMA stuff which is not enabled on e200 and 3g
22:26:45gevaertsmax131: I'd say the sd driver
22:27:05gevaertsAlthough usb isn't impossible
22:27:49jhMikeSBuschel: r28976
22:28:16Buschelpretty up-to-date
22:28:24max131is there maybe an easy way to trace back and see when the last change to the sd driver was made?
22:28:34 Join fml [0] (
22:28:35TheSevensound.c:199 looks like an apparent bug
22:29:16fmlgevaerts: many kudos for fixing plugins.c. You did it in a much simpler and elegant way than I would have done it!
22:29:55 Quit factor_ (Read error: Connection reset by peer)
22:30:01gevaertsfml: I'm still thinking the whole thing ought to be done in a simpler way, but this was the simplest change on its own I think
22:30:14 Join factor [0] (~factor@
22:30:52fmlgevaerts: yes, if all would be done correctly and consistently from the beginning, the fix wouldn't be needed at all
22:31:05max131gevaerts: is it sd-as3525.c that you are talking about?
22:31:28*TheSeven considers sound.c an awful unflexible ifdef hell
22:31:35gevaertsmax131: probably. I personally never worked on the AMS players though, so I can't really help much
22:31:51jhMikeSTheSeven: That's line is certainly odd to look at
22:31:58max131gevaerts: who should i try and contact then?
22:32:22TheSevenjhMikeS: i missed a set of parantheses when looking at it for the first time
22:32:26gevaertsmax131: svn log on that file should tell you :)
22:32:33kugeldidn't wodz work on sound.c cleanup?
22:32:58 Quit fml (Client Quit)
22:33:09CIA-7New commit by 03wodz (r28998): HD300 - calibrate discharge curve
22:33:25TheSeven"The codecs listed use HW tone controls but don't have suitable prescaler functionality, so we let the prescaler stay at 0 for these, unless SW tone controls are in use. This is to avoid needing the SW DSP just for the prescaling." - now what the hell is that meant to tell me?
22:34:12TheSevenit doesn't seem to match the code on the first glance, as it's an exclusion, not inclusion, list
22:34:40max131gevaerts: ok, thanks
22:35:08TheSevendo codecs that use hardware tone controls define HAVE_SW_TONE_CONTROLS?
22:35:19TheSevenapparently they do, judging from that code
22:35:25CIA-7r28998 build result: All green
22:37:15jhMikeSTheSeven: I'm not sure what this "suitable prescaler functionality" is. Sometimes it needs to be manually calculated but the hw can still prescale.
22:40:06jhMikeSex: if wm8721 has tone controls off, then it shouldn't call prescaler code, since it doesn't need it without tone controls
22:40:23jhMikeSit doesn't have it's own anyway
22:41:55 Quit advcomp2019_ (Quit: IRC is just multiplayer notepad)
22:42:33jhMikeSsound.c is messy :\
22:44:15leavittxhi! i run a simulator of cowon d2 (also tried x5 - same thing) and get segfault when i choose for example plugins->games->xobox and then Properties. this is strace output:
22:48:50perfectdrugwodz: you may be interested in
22:49:42 Quit krabador (Quit: Sto andando via)
22:50:40wodzperfectdrug: awesome
22:51:57perfectdrugthe hard work was already done otherwise I wouldn't have had the time for it
22:54:35bluebrotheranyone familiar with perl and can have a look at
22:54:51leavittxwhat generally could be a reason of simulator segfault?
22:55:23TheSevengdb will tell you :)
22:57:16TheSevenwhat could cause this?
22:57:19TheSeven/data/rockbox-trunk/apps/settings_list.c:1423: error: âLANG_BASS_CUTOFFâ undeclared here (not in a function)
22:57:27CIA-7New commit by 03wodz (r28999): Add thumbnail for HD300
22:58:14wodzTheSeven: this string is not defined in lang file for this target
22:58:23TheSevenyeah, but why?
22:59:15*bluebrother wonders if he should just commit this. But: do I need to update the revision number?
22:59:32wodzB4gder: ping
22:59:39jhMikeSThat string might not be using features properly
22:59:50leavittxTheSeven: i don't believe that only i have this segfault... at least my friend have it too
22:59:53B4gderI'm here
23:00:41jhMikeSnope, only defined for ipodvideo
23:00:44wodzB4gder: May I ask for www update?
23:00:45gevaertsbluebrother: that's probably best
23:00:53TheSevenjhMikeS: where do i need to change that?
23:00:57bluebrotherB4gder: do I need to change the revision number when changing to fix cpu core number detection on OS X?
23:01:10TheSevenB4gder: do you happen to know where Zagor is hiding or when he'll be back?
23:01:13B4gderwodz: sure, done!
23:01:17wodzB4gder: I just commited thumbnail for HD300 (thanks to perfectdrug's work)
23:01:21bluebrother(see −− any comments on that? I'm not that much familiar with perl, but it works for me :)
23:01:27gevaertsbluebrother: if you don't change it, existing clients can't be told to upgrade
23:01:31B4gderbluebrother: if you want all clients to upgrade, then yes
23:01:56wodzB4gder: thx
23:01:56 Quit max131 (Quit: CGI:IRC (EOF))
23:02:13bluebrothergevaerts: sure, but do I want that? It's only relevant for OS X, and since the number of cores gets reported as 0 right now on OS X it doesn't work anyway.
23:02:19B4gderTheSeven: yes I do know where he is physically but I can't tell when he'll be around in here
23:02:44jhMikeSTheSeven: or apps/features.txt if you want fix the feature part
23:02:48bluebrotherhmm, maybe it's best to bump the revision number, wait for clients to update and see if I brok the build system :)
23:03:00perfectdrugwodz: thanks ( B4gder too) do you commit the manual part aswell?
23:03:39wodzperfectdrug: When I start work on manual for HD300 I'll commit this
23:03:55 Quit factor (Read error: Connection reset by peer)
23:04:24 Join factor [0] (~factor@
23:04:26bluebrotherB4gder: any concerns about the change?
23:04:43B4gderI haven't looked at it, I'm a bit busy elsewhere atm
23:04:46 Quit Xerion (Quit: )
23:05:39TheSevenjhMikeS: should I just duplicate that line?
23:08:40 Quit captainkewll (Quit: Page closed)
23:12:28bluebrotherhmm, if I want the build clients to update I also need to bump $minimumversion in, right?
23:13:38gevaertsYes. I'm not sure if it's a good idea for anyone but a server admin to actually do that though
23:14:09bluebrotherhmm. I guess I'll keep the current revision number then.
23:14:50bluebrothersince that change isn't required by any client running on linux. And anyone setting up a client on OS X should (hopefully) get the most recent scripts from SVN anyway.
23:16:48BuscheljhMikeS: a question -> what happens if the codec thread gets higher priority as the buffering thread, if rebuffering is occuring?
23:18:31 Quit Keripo (Quit: Leaving.)
23:18:37jhMikeSBuschel: It would cause codec to be preferred over buffering if both have to run, but if the codec is waiting for buffering, the buffering thread will run
23:21:15 Quit perfectdrug (Quit: Leaving.)
23:22:00jhMikeSit would only be a problem if the codec can't regain ground and fill the PCM buffer. if it can't, then CPU resources are being overtapped
23:22:24 Quit wodz (Quit: Leaving)
23:23:03 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
23:25:22 Quit komputes (Remote host closed the connection)
23:26:22CIA-7New commit by 03bluebrother (r29000): Make probecores work on OS X. ...
23:31:47 Join simonrvn_ [0] (simon@
23:32:35 Quit simonrvn (Ping timeout: 240 seconds)
23:32:35 Nick simonrvn_ is now known as simonrvn (simon@
23:45:12TheSevenjhMikeS: hm, the prescaler thing doesn't seem to work quite right
23:45:21TheSevenstarting at bass +3dB i get quite noticable clipping
23:45:39TheSeventhe overall volume is constant now though
23:49:19TheSevenif i don't offer a hw prescaler, things seem to be a bit better, but there's still noticable clipping at +6dB
23:51:41jhMikeSI'd be hard pressed to know why for your codec, I've only integrated that for wolfons. double-check scaling of db units?
23:52:29 Quit kugel (Remote host closed the connection)
23:52:46 Quit Buschel (Quit: ChatZilla 0.9.86 [Firefox 3.6.13/20101203075014])
23:57:31TheSevenjhMikeS: is my assumption correct, that it *will* use the software DSP, even if it needs it just for precut gain? (in contrast to the comment, which seems to be referring to compiling it in, not using it)
23:58:58 Quit kadoban (Read error: Operation timed out)

Previous day | Next day